One of the ways that machine learning algorithms can help control the spread of an infectious disease is by building models that predict who is likely to become infected making them good candidates for preemptive interventions. In this work we ask: can we build reliable infection prediction models when the observed data is collected under limited, and biased testing that prioritizes testing symptomatic individuals? Our analysis suggests that when the infection is highly transmissible, incomplete testing might be sufficient to achieve good out-of-sample prediction error. Guided by this insight, we develop an algorithm that predicts infections, and show that it outperforms baselines on simulated data. We apply our model to data from a large hospital to predict Clostridioides difficile infections; a communicable disease that is characterized by both symptomatically infected and asymptomatic (i.e., untested) carriers. Using a proxy instead of the unobserved untested-infected state, we show that our model outperforms benchmarks in predicting infections.

Mobile robot tipover is a concern as it can create dangerous situations for operators and bystanders, cause collateral damage to the surrounding environment, and result in an aborted mission. Algorithms have been developed by others to assess the stability of the robot, and many of these algorithms have been demonstrated using simulated data. In order to verify that these algorithms accurately match real-world behavior, we have collected data of a mobile robot tipping over and then compared this data to the stability measures provided by three algorithms: Zero-Moment Point (ZMP), Force-Angle stability measure (FA), and Moment-Height Stability measure (MHS). A small mobile robot platform based on the iRobot PackBot drove a course including ramps and obstacles; an IMU and GPS provided inertial and positional data for the algorithms, and the actual tipover event is determined from video footage of the tests. The average normalized measure at tipover event initiation was found to be 0.665 for ZMP, -0.094 for FA, and 0.023 for MHS, where a value of 1 corresponds to resting stability. Standard deviations were 0.38, 0.84, and 0.67, respectively. The measures show a significant amount of noise, which is likely due to the vibrations caused by movement of the tracks and could be reduced by employing additional filtering during data collection. The preliminary real-world data validates these tipover algorithms as able to assess robot stability, and they can be used as part of a tipover avoidance system.